Remora - ethernet NVEM cnc board

www.nxp.com/products/processors-and-micr...-m7-core:i.MX-RT1050

How is the rt1059 currently configured for boot? Page 197 of the " i.MX RT1050 Processor Reference Manual" 

Also found in this app note  www.nxp.com/docs/en/application-note/AN12107.pdf
"The boot ROM uses the state of the BOOT_MODE register and eFUSEs to determine the boot device. For development purposes, the eFUSEs used to determine the boot device may be overridden using the GPIO pin inputs."

I'm guessing the OEM has set the eFuses to boot from the QSPI flash. This should not stop a SWD connection (I hope).

I've just got a DAPlink buid environment working and compiled the DAPlink firmware for the ST-Link v2. First step is to check the standard stm32f103xb_bl firmware is working before trying to modify the USB vendor and product ID's. Fingers crossed

SWCLK/SWDIO/GND must be in same order for different versions of the PCB. Because it's logically right But I can't figure it out right now, because a solder station isn't available for me now. I tried to change clk/io pins while using the DAPLink (ST-LINK v2 clone), but with no luck.

Yes, I think for the OEMs eFuses (or values softly set by a bootloader) must be set for the boot via SPI flash. It's also logically right

One more option for the quick firmware download - use a cheap SPI flash programmers like this aliexpress.ru/item/32793476447.html. If a built-in bootloader allows to use any SPI flash firmware (without any checks), this method will be the easiest way for us and other users.

SPI programmer is probably the way to go. Just have to verify that there are no problems with it working in circuit.

In some cases SPI flash can be not readable/writable when it soldered to the PCB. In this cases we must to solder it out. But I have a feeling that SPI flash is encrypted and the bootloader stores the keys for the decryption. Bootloader must be not readable (due to copyright purposes). So going by the SPI flash way can be a bad idea.

By the way, we have a button on the PCB. I guess it makes something with a boot sequence. When the ST-Link is connected and the button is pressed on boot, additional LED is lighting. In other cases this LED is off (see the photos).

A magic "IC" chip is also here (see the photos).

Why the risk with these boards is them actually having the flash "encrypted and the bootloader stores the keys for the decryption" if not, yet but easily changed in the future by the maker. Can you change the keys or remove them after on the RT1050? Can you return the rt1050 to a blank slate after they have set the keys? We don't care about cracking their binary firmware, just replacing it.Do you really want to spend this much time on a port if other lower cost boards are available?

Interesting to see the NVEM v2.1 also has the push button and voltage monitoring circuit and that there appears to be a boot sequence.

I feel a bit bad that people went out an purchased an NVEM thinking that they could convert it for use with LinuxCNC.

Will give this a good go to see if we can connect to the MCU, as you said we don't really care about the existing firmware, just being able to replace it.

scotta wrote:

I feel a bit bad that people went out an purchased an NVEM thinking that they could convert it for use with LinuxCNC.
 

 

Just thinking out aloud here, is there a processor that is more suited to Remora ? I don't claim to be an Electronic Engineer but I have a bit of experience in PCB design and layout, I have built my own servo drives for several of my own machines using Elm Chan firmware, first incarnation worked for 9 years without failure.

From what I've seen these controllers aren't particularly hi-tech, buffering, optos, packaged fairly nicely. If there is a processor that could be used that is readily available at the moment I would be happy to do a prototype layout and get PCB's built for testing. I've heard the STM blue pills and the like are very cheap but also there are a lot of fakes.

We would have to keep in mind being an open source project that people can replicate at home, so possibly something like an STM blue pill/esp32 whatever that can be socketed onto a through hole board without SMD tooling requred.


 

"NVEM + LinuxCNC" project it's a hobby for me now. I made my own pair of a firmware/driver for the old NVEM (STM32F207). And my work (and scotta's work too) can be moved into the trashcan soon (because of RT1052). My previous project (Orange Pi + LinuxCNC, "allwincnc") was archived because new Linux kernel/uboot pair breaks my work on the built-in CPU's firmware.

So I decided to force the development of my own controller for the LinuxCNC. I planning to use my code (partly) from both of those projects. The PCB will be a simple shield for the cheap modules. See the photos below. Some parts will be replaced to another ones for the simplicity. Planning to create a new topic soon.

The problem with the OrangePi's is availability, that is the only reason why we haven't built a good Linux and u-boot for them.

forum.linuxcnc.org/18-computer/44828-rem...ard?start=480#252695 has a ARM SOC for Linux and LCNC plus a STM32 for Remora for ~$60 usd.

shop.lerdge.com/products/adapter-module  for the stepper driver headers and external stepper drives

@epineh a few STM32's have been ported to Remora. The problem is availability of them since they often have one year lead times or only pop up sporadically.

I design single board computers and my issues are which parts to pick to be sure they are available for the next year or two. I am happy to even design a board for the Chinese to copy and produce cheaply.

forum.linuxcnc.org/18-computer/46077-pot...ters?start=20#251959 is also of interest